The Google Coral USB Accelerator brings real-time inference to your Pi 4 and many other computers!
Artificial Intelligence / Machine Learning for everyone: Google has connected the Coral USB Accelerator, a powerful special chip (TPU, Tensor Processing Unit) to a USB 3 interface - so Tensor Flow Lite models can be used for inference quickly and in an energy-saving way. A particular advantage of this solution: Your data remains local. This helps with latency, and of course with data protection!
Google increasingly uses artificial intelligence (AI) and machine learning (ML) to realize its services. To this end, it has developed specialized processors called TPU (''tensor processing unit'') for its data centers, which can execute algorithms faster and more energy-efficiently using the TensorFlow framework. For example, Google Maps is enhanced by street signs recorded by Street View, which are analyzed using a neural network based on TensorFlow. The highlight: TensorFlow can be easily programmed in Python.
Google launches a USB 3 stick with the Edge TPU, which supports the TensorFlow Lite framework. The Edge TPU can perform up to 4 trillion calculations per second with only 2 W consumption.
Perfect in combination with the Pi 4!
With the help of the Google Coral Edge TPU, inference can be performed up to 20 times faster on the MobileNet v2 model than on the ''naked'' Pi 4. This allows real-time detection in video streams at over 50 fps, which would not be possible with the Pi 4 without an accelerator.
Thanks to Python and many examples on the Internet around TensorFlow, you can get into the topic of artificial intelligence and machine learning with the Google Coral USB Accelerator easily and with style.
Technical data Coral USB Accelerator
• Google Edge TPU ML accelerator coprocessor
• USB 3.0 (USB 3.1 Gen 1) Type C socket
• supports Linux, Mac and Windows on the host system
• Power consumption up to 900 mA peak @ 5 V
• recommended ambient temperature max. 25°C (maximum computing power)
• maximum ambient temperature is 35°C (reduced clock frequency)
• Dimensions (LxWxH): 65 x 30 x 8 mm
Requirements for the host system
• Linux Debian 6.0 or higher, or a derivative thereof (e.g. Ubuntu 10.0+, Raspbian)
• System Architecture: x86-64, ARMv7 (32-bit) or ARMv8 (64-bit)
• macOS 10.15 installed with either MacPorts or Homebrew
• Windows 10
• a free USB port (should be USB 3 for best performance)
• Python 3.5, 3.6 or 3.7
Scope of delivery Google Coral USB Accelerator
• USB Accelerator
• USB 3 cable
Notes
• the USB flash drive can become very hot during operation, which can cause burns - please wait until it has cooled down before touching it!
• Google and we are not responsible for any damage if the device is operated...
